A category-by-category reference for the findings behind your score — start with the overview, then go deep on whichever category is dragging your report down.
How FlowScan v2.0 turns an exported n8n or Make workflow into a 0–100 score: the seven categories, how points are deducted, the three hard caps, the Cost Exposure Score, the Controls Matrix, and the language rules every finding follows.
What the Reliability category checks — error handling, idempotency, branching integrity, chain failure risk, and structural hygiene — with examples, score impact, and recommended fixes.
What the Cost Risk category checks — loop multipliers, oversized prompts, missing token caps, and chatty agents — with examples, score impact, and recommended fixes.
What the AI Governance category checks — write-capable agent tools, prompt-injection-shaped paths, and shadow AI calls — with examples, score impact, and recommended fixes.
What the Observability category checks — audit logs and execution summaries for consequential writes — with examples, score impact, and recommended fixes.
What the Security category checks — unauthenticated webhooks and hardcoded secrets — with examples, score impact, and recommended fixes.
What the Naming category checks — generic default names and duplicate node names — with examples, score impact, and recommended fixes. Low severity, still worth fixing.
What the Complexity category checks — node count and path depth thresholds — with examples, score impact, and recommended fixes for splitting an overgrown workflow.